These experiments are designed to determine whether adolescents at risk for substance abuse (SA) exhibit impaired cognitive performance on a task that reveals a deficit in performance by adults with a history of SA. Seventy-four adolescents (28 healthy and 46 ADHD; age at study entry: 12.7 + 0.7; age at last follow-up: 16.1 + 1.7) participated in an ongoing longitudinal study of predictors of substance use. The substances most commonly reported were nicotine, alcohol, and marijuana. Aggression, conduct problems, hyperactivity, impulsivity, inattention, anxiety/depression, social difficulties, and somatic complaints were assessed at study entry, and tested as predictors for later substance use. Four years into the study, 34 adolescents had not used any substances, 39 had experimented with at least one substance, and 26 with more than one substance. Psychiatric diagnoses (healthy, ADHD, ADHD & CD, and ADHD & depression/anxiety) did not influence reports of substance use. Distinct behavioral measures collected at study entry predicted use of different substances. Hyperactivity had the greatest association with tobacco smoking. Aggression and fewer social problems were associated with marijuana use. Severity of drug use, indexed by the number of substances used, was predicted by somatic complaints, hyperactivity, and fewer social problems. This four-year longitudinal study captured onset of substance use, not abuse. Behavioral predictors differed with type of substance used. These behavioral characteristics may raise suspicion among pediatricians for enhanced risk for substance use initiation. Self-reports and parental reports of difficulties associated with executive function (e.g., planning, organization and initiation) are significantly correlated with an external locus of control in 37 adolescent boys (aged 12-18, with and without ADHD) enrolled in this study. An external locus of control, which has been associated with adaptation and emotional difficulties such as depression and anxiety, is also sensitive to the onset of substance abuse. Follow-up studies of these participants will test the association of perceived control with dysfunctional adaptations such as substance abuse. Using data acquired from adolescents enrolled in this study, we evaluated the inter-rater reliability, internal consistency and/or temporal reliability of a number of the test instruments. We assessed the internal consistency and temporal stability of Conners? Continuous Performance Test (CCPT) as a clinical tool for the assessment of attention-deficit/hyperactivity disorder (ADHD) the CCPT in a sample of 51 adolescents, (19 with ADHD, 8 with ADHD and another disorder, and 24 controls). The CCPT has adequate split-half reliability, and performance was found to be moderately correlated over the span of 1.3 years. ADHD participants tended to perform worse than controls during the first, but not second, CCPT administration. The CCPT was found to be an insensitive indicator of diagnostic status and, while several CCPT performance measures correlated with parent ratings of ADHD symptomatology, the pattern of these correlations was largely independent of the interpretive guidelines provided in the test manual. We conclude that the poor diagnostic validity and the unclear functional relevance of CCPT scores might confound clinical interpretation of the CCPT. We assessed the inter-rater agreement and internal consistency of the Rey Complex Figure Test (RCFT) in a sample of adolescents enrolled in this study. We found each of the RCFT outcome measures to be associated with clinically acceptable inter-rater reliability estimates. Additionally, we found the copy and recall accuracy scores to be associated with clinically acceptable internal consistency. The organizational scoring method we evaluated, while associated with internal consistency that is acceptable for research purposes, was found to be somewhat below estimates associated with clinical acceptability. We assessed the short- and long-term stability of the Stroop Color and Word Test, a commercially-available version of Stroop?s original task in 87 adolescents. Participants were tested an average of one month apart at baseline and twice more at approximately 15 month intervals. We ffound that all of the commonly used performance indices are associated with practice effects and that the three primary scores, but not the formula-derived interference score, were associated with acceptable reliability.